Imago 1.0 Beta

[Fedor Zentsev]

Hello all,

It is a pleasure to announce that we have released our new product:
Imago - open-source library for 2D chemical structure image
recognition.

Imago 1.0 Beta is available on http://scitouch.net/imago
as well as Ego - Java GUI application and Alter-Ego - command-line
utility, both built on Imago library.

We will be glad to hear feedback about bugs, new features, etc.

With best regards,

Fedor Zentsev
SciTouch LLC

[Igor Filippov]

First - congratulations on your 1.0 release!

I would be very interested to see the recognition rates for Imago.
There are publicly available validation sets of images, for example
ChemInfty set derived from Japan Patent Office images
http://www.iapr-tc11.org/mediawiki/index.php/ChemInfty_-_Chemical_Structure_GT
(I recommend "focused" subset)
and USPTO set:
https://sourceforge.net/apps/mediawiki/osra/index.php?title=Validation

both sets include ground truth mol files.
Also, there are validation images from CLiDE project, though they do
not
have the ground truth files available:
http://www.simbiosys.ca/clide/validation.html

I believe I've discussed the validation sets with Dmitry on his blog
at one point -
As I discussed with Dmitry on his blog
http://shmat-razum.blogspot.com/2010/07/blog-post.html

Best regards,
Igor Filippov
OSRA developer

On Aug 1, 7:21 pm, Fedor Zentsev <fzent...@scitouch.net> wrote:
> Hello all,
>
> It is a pleasure to announce that we have released our new product:
> Imago - open-source library for 2D chemical structure image
> recognition.
>

> Imago 1.0 Beta is available onhttp://scitouch.net/imago

[fzen...@scitouch.net]

Hello Igor!

Thank you for your congratulations and information about
validation sets you have provided.

At the moment, unfortunately, Imago do not posses the ability to
expand superatom labels.
This is the reason why we are not able to provide accurate recognition
rate now:
recognized molecule doesn`t match with corresponding molfile from the
ground truth set,
however extracted information is correct.

With best regards,
Fedor Zentsev

Scitouch LLC

On 21 сен, 23:39, Igor Filippov <igor.v.filip...@gmail.com> wrote:
> First - congratulations on your 1.0 release!
>
> I would be very interested to see the recognition rates for Imago.
> There are publicly available validation sets of images, for example

> ChemInfty set derived from Japan Patent Office imageshttp://www.iapr-tc11.org/mediawiki/index.php/ChemInfty_-_Chemical_Str...

> (I recommend "focused" subset)
> and USPTO set:https://sourceforge.net/apps/mediawiki/osra/index.php?title=Validation
>
> both sets include ground truth mol files.
> Also, there are validation images from CLiDE project, though they do
> not
> have the ground truth files available:http://www.simbiosys.ca/clide/validation.html
>
> I believe I've discussed the validation sets with Dmitry on his blog
> at one point -

> As I discussed with Dmitry on his bloghttp://shmat-razum.blogspot.com/2010/07/blog-post.html

[Chupakhin Vladimir]

Браво! Вы просто ворвались на рынок хемоифнорматики!
Мои поздравления!